A Geothermal GIS for Nevada'''

1
A Simplification of Weights of Evidence using a
Density Function and Fuzzy Distributions, using
Geothermal Systems in Nevada as an example.
Using a Geothermal GIS started by Mark Mihalasky
Mark F. Coolbaugh, Richard L. Bedell Great Basin
Center for Geothermal Energy, University of
Nevada, Reno USA and GeoCorp
Funding To the Great Basin Center for Geothermal
Energy, through the cooperative grant
DE-FG07-02ID14311 administered by DOE, Idaho
Operations Office and the Idaho National
Engineering and Environmental Laboratory.
Beowawe, photo Jack Quade
2
Newberry Crater
Geothermal Systems in Nevada Great Basin, USA
Big Southern Butte
China Hat
Borax Lake
Medicine Lake
Roosevelt/ Cove Fort
Peak weights of evidence contrast 6.6
Mammoth
Coso
Boundary of Great Basin
Current capacity 600 MWe
3



Young Faults (NBMG/USGS)
Young Volcanics (USGS)
Boron, Groundwater (NWIS)
Heat flux (SMU)
Binary WofE Statistics



Quakes (NV Seismo Lab)
Depth to Groundwater (NWIS)
PZ Carbonates (NBMG)
4
Geothermal system density, / km2
Heat Flow, mW / m2
From Wisian, Blackwell, and Richards 24th
Geothermal Reservoir Engineering Workshop,
Stanford, CA., 1999, p. 219-226.
5
Probability Modeling with continuous data
distributions can be reduced to one ratio
(density).
For one interval of data
Ni / Ai
of deposits intersecting the data interval i /
Area of pattern i
Normalize this ratio and convert to a unitless
number by dividing by the overall study area
deposit density
Ni / Ai
Nt / At
area of the pattern for data interval i / the
total study area
Steamboat Springs geyser, photo Don Hudson
6
Density Function Model
DF (Ni/Nt) / (Ai/At) Ni number of
deposits associated with pattern i Nt total
number of deposits Ai area of the pattern
At total area of the study
When expressed in log form, this is similar to
WofE when 1) area of deposits in a pattern is
small compared to the total area of that
pattern, and 2) total area of deposits is small
compared to total study area.
Steamboat Springs geyser, photo Don Hudson
7
Similarity between Density Function and Multiple
Weights of Evidence, for Heat Flux and Geothermal
Systems in Nevada
59 training points 9 km2 per training point Study
Area 286,000 km2
lt 0.5 difference Difference is lt 1 of ? of W
8
Fitting of a Smoothed Density Function for Heat
Flux
9
Density Function for Earthquakes
10
Binary WofE
Density Function
11
(No Transcript)
12
Binary WofE
Density Function
13
CONCLUSIONS Density Function

• When dealing with rare occurrences,
  weights-of-evidence can be closely approximated
  by a density function.
• A categorically-based density function can help
  constrain continuous fuzzy membership functions.
• Smoothing of multiple weights (either statistical
  or expert-based) can provide more uniform vectors
  towards favourable terrain.

14
The End
Reno, NV.
15
Buffalo Valley Anomaly is based on1) Humboldt
lineament, 2) very young basaltic volcanism,
3) Battle Mountain heat flow high, and 4)
elevated crustal extension as measured by GPS.1
and 4 were not used in the model. (Purple circles
are geothermal power plants)
Binary WofE
Buffalo Valley
Steamboat Springs
Humboldt Structural Zone
Desert Peak/Soda Lake
Density Function
16
Binary WofE
Density Function
17
Example from Kriging Variograms
Taken from M. David, 1982, Geostatistical Ore
Reserve Estimation
18
Gravity and magnetics showing original data
with corresponding histogram and data intersected
by 1 km buffers of geothermal well location.
(a) Bouguer Gravity, (b) Bouguer Gravity
intersected with 1km buffers around geothermal
wells.
19
Original Data.Original Data Intersected
with Exploration Target.Cumulative Frequency
Curves- relatively insensitive to data
distributions.- area between curves defines
unbiased difference or
potential contribution.- Normalize
20
Difference between Weights of Evidence
Density Function Rankings
21
(No Transcript)
22
Binary WofE vs.
Density Function